Value creation with passive socially responsible exchange-traded funds

(1)

Value creation with passive socially responsible exchange-traded funds

Vaasa 2021

School of Accounting and Finance Master’s thesis in the Finance Programme

(2)

UNIVERSITY OF VAASA

School of Accounting and Finance

Author: Jesse Järvinen

Topic of the thesis: Value creation with passive socially responsible exchange-traded funds

Degree: Master’s Programme in Finance Master’s Programme: Finance

Supervisor: Nebojsa Dimic Year of completing the

thesis:

2021 Number of

pages: 114

ABSTRACT

:

Using a unique data set of 121 U.S. equity ETFs from January 2010 to December 2020, this thesis studies whether passive socially responsible exchange-traded funds overperform their passive counterpart that does no socially responsible screening.

Further, this thesis examines what are the strategies (inclusion or exclusion) and attributes (ESG and product-related) that drive the performance results.

Socially Responsible Investing (SRI) has grown substantially during recent years in asset size and as products that incorporate the idea of “doing good while doing well.” Socially responsible investing seeks to deliver returns while evaluating the long-term impact of a company’s business policies on society and on the environment. Mutually exclusive, the exchange-traded funds (ETFs) are growing substantially in assets under management and product offering. While expensive active asset management steers investors towards passive and less expensive alternatives, the ETFs offer a transparent and cost-efficient way to practice different investing styles like the SRI. The SRI ETFs are a recent financial innovation, academic interest is emerging, and this thesis intends to fill it by examining the existing literature of SRI and ETFs individually and combined.

The empirical part of this thesis provides answers for investors considering SRI ETFs. An equally weighted SRI ETF portfolio underperforms the passive counterpart of an equally weighted portfolio of passive S&P500 ETFs significantly on the longest sample period of January 2010 to December 2020. However, the periods after January 2015 offer distinctive results as the SRI ETF portfolio overperforms the counterpart. It seems to be relating to the development of SRI ETFs that substantially grew in asset size and products after 2015. Furthermore, it seems to be the ETFs using Environmental Inclusion (positive screening) as a strategy that drives the SRI ETF sample group abnormal returns. The ETFs using Environmental Inclusion overperform the other screening strategies statistically and economically significantly.

This thesis provides evidence that financial performance does not consistently exclude sustainable performance and that by choosing a passive socially responsible ETF, investors can “do good while doing well.”

KEY WORDS: Passive asset management, socially responsible investing (SRI), exchange-traded funds (ETFs), Environmental, social governance, ESG, Modern Portfolio Theory

(3)

List of Figures

Figure 1. U.S. Equity fund flows between passive and active funds. 14

Figure 2. The minimum-variance frontier. 20

Figure 3. Asset-weighted average fees for funds as the end of 2018. 24 Figure 4. Actively managed U.S. equity funds outperformed by benchmarks. 26 Figure 5. Assets under management of global ETFs from 2003 to 2019. 28

Figure 6. The ETF architecture. 31

Figure 7. Sustainable and Responsible Investing in the United States 1995-2018. 44

Figure 8. Assets Under PRI Management. 45

Figure 9. Examples of the ESG factors. 52

Figure 10. Top Specific ESG Criteria for Money Managers 2018. 53

(6)

List of Tables

Table 1. Minimum diversification requirements for RICs and UCITS. 35 Table 2. The six Principles of Responsible Investment. 47

Table 3. Responsible investing strategies. 48

Table 4. Summary statistics of the SRI ETFs. 73

Table 5. Descriptive statistics. 75

Table 6. Regression results from the CAPM. 84

Table 7. The CAPM factor loadings for the full sample period. 85 Table 8. Regression results from the three-factor model. 86 Table 9. The three-factor model factor loadings for the full sample period. 88 Table 10. Regression results from the four-factor model. 89 Table 11. The four-factor model factor loadings for the full sample period. 91 Table 12. Regression results with the four-factor model after controlling for the fees. 93 Table 13. Regression results for the different screening criteria and ESG attributes. 95 Table 14. Factor loadings for all of the different screening strategies. 98

(7)

Acronyms

AUM Assets Under Management

CAPM Capital Asset Pricing Model

CRSP Central for Research in Security Prices CSR Corporate Social Responsibility

CFP Corporate Financial Performance CSP Corporate Social Performance ESG Environmental, Social, Governance

ETF Exchange Traded Fund

EU The European Union

NAV Net Asset Value

NYSE New York Stock Exchange

PRI Principles of Responsible Investing RIC Regulated Investment Company

SEC The Securities and Exchange Commission

SIB Social Impact Bond

SIF The Forum for Sustainable and Responsible Investment SRI Socially Responsible Investing

TER Total Expense Ratio

TNA Total Net Assets

UCITS Undertakings for Collective Investment in Transferrable Securities

UK The United Kingdom

UN The United Nations

US The United States

(8)

1 Introduction

The fundamental function of investing is the expectation of obtaining an additional profit or income in the future. The history of financial markets and investing is everlasting and encompasses many different trends as well as many famous practices. Among recent financial market trends, socially responsible investing and passive asset management have turned from a minor niche to mainstream style investing. Furthermore, these two phenomena are emerging different strategies and instruments to practice them (UN PRI, 2020.)

Interest in sustainable development is recognized worldwide, and it continues to be one of the most talked-about topics in society nowadays. Sustainable investing generally refers to acronyms like socially responsible investing (SRI) in addition to consideration of environmental, social, and corporate governance (ESG) issues. Socially responsible investing seeks to deliver returns while evaluating the long-term impact of a company’s business policies on society and on the environment. Investors are not only seeking financial returns but additionally they are also promoting environmental and social objectives. When interest and awareness towards these issues arise among investors and increasing demand from legislators, they are corollary resulting in companies to integrate socially responsible and sustainable matters into their business processes (Renneboog, Horst & Zhang, 2008). Whatever the effect is on companies, previous evidence (Waddock, Bodwell & Graves, 2002; Konar & Cohen, 2001) suggests that environmental, social, and governance issues impact companies’ value, and managers can no longer ignore this.

It was not long when this niche investment style started developing, and the asset managers began to practice SRI and incorporate the ESG issues into their investment decision-making process and investment analysis. Socially Responsible Investing (SRI) has grown substantially during recent years in asset size and as products that incorporate the idea of “doing good while doing well.” Socially responsible funds began excluding stocks that deal with tobacco, alcohol, gambling, and fossil resources, usually referred to

(9)

as “sin-stocks.” Further, the SRI funds use different screening strategies like inclusion (positive screening), where the asset manager selects stocks with the best ESG attributes, and exclusion (negative screening), where the asset manager rules out the stocks with the worst ESG attributes (Nofsinger and Varma, 2014).

Simultaneously, another relatively new financial phenomenon is gaining popularity among academic researchers and investors. Passive asset management and index investing have been grounding reasons for a relatively new investment innovation: The Exchange Traded Funds (ETFs). As French (2008) research suggests, active asset management costs are comparatively high, and it steers investors towards passive asset management strategies. The majority of the ETFs typically seek to follow a specific index’s performance, as an index mutual fund does, but ETFs vary in many crucial ways from the mutual funds. More of the differences are described later in this thesis. Passive asset management results in fewer transactions, leading to more cost-efficiency. Also, it avoids the risk creation through untimely or mistakenly treated actively managed assessments. The ETFs are growing substantially in assets, diversity, and market significance after their appearance in the financial markets. Among studies, ETFs have evidence of their efficiency in the financial market, and investors are well aware of these possibilities. Up to this date, there are several ETFs available to invest in a socially responsible manner and this study identifies 121 SRI ETFs during the sample period.

Riedl and Smeets (2017) shed light on the question of why investors hold SRI funds. Their study finds that intrinsic social preferences and social signaling are important elements for investors to hold SRI funds. Financial motives also play a role but to a lesser extent since investors are willing to pay higher management fees, and they expect SRI funds to underperform the conventional counterpart. Furthermore, investors with a longer investment horizon are willing to choose SRI funds.

These two trends have the potential for dramatically reshaping the broader investment landscape, like their year-to-year growths are proven to be nonetheless remarkable.

Socially responsible investing remains influential in the investors’ minds and legislators’

(10)

minds forcing companies to consider environmental and social issues. Mutually exclusive, investor’s interest in cost-efficient and easily distributed investment solutions is only likely to validate ETFs market growth. While the SRI ETFs usually have higher sustainability rankings than the conventional counterparts, they still have higher costs due to strict screening strategies.

This thesis will merge these two phenomena by examining the existing literature of these two subjects. Using a unique dataset of U.S. equity SRI exchange-traded funds for the period of January 2010 to December 2020, this thesis investigates passive SRI ETFs’

performance over a passive ETF counterpart that follows the S&P500 index. Further, this study empirically tests what are the ESG attributes and strategies in the SRI ETFs that create value for investors.

1.1 Hypothesis development and purpose of the study

As socially responsible ETFs being a current topic, academic interest is emerging, and this thesis intends to fill it. This study’s primary purpose is to examine whether socially responsible ETFs generate alpha for investors over the conventional counterpart and what are the strategies (inclusion or exclusion) and attributes (ESG and product-related) that drive the results. There are many studies examining the effect of socially responsible investing on financial performance (e.g., Bauer, Koedijk & Otten, 2005; Chang, Nelson &

White, 2012; Nofsinger & Varma, 2014), yet none of them focus on ETFs. This thesis aims to provide answers for investors considering ETFs as well as socially responsible investing issues. It offers a comprehensive review of passive and active asset management, ETFs, socially responsible investing, and why and how these topics are generating value for investors.

Because socially responsible ETFs are a relatively recent financial instrument, there is not much research nor data on socially responsible ETFs nor their performance. However, it has gained much attention from individual investors. While existing literature is mainly

(11)

focusing on socially responsible mutual funds, this study examines ETFs and their potential for creating financial performance with SRI. Indeed, year-to-year growths of both ETFs and SRI give evidence that there is something the investors value in these investments. There is much evidence that investors pay a price for investing in a socially responsible way (Girard, Rahman & Stone, 2007; Renneboog et al., 2008; Humphrey and Tan, 2014). Alternatively, there is evidence that integrating SRI into financial investments they can outperform the conventional counterpart (Derwall et al., 2005; Kempf and Osthoff, 2007; Bebchuck et al., 2009). Concluding on all of these examples, the first hypothesis in this thesis is the following:

H1: Passively managed socially responsible ETFs are not consistently losing to passive ETF counterparts with no SRI screening.

Another objective is to examine whether the different ESG attributes create value differently and that the different screening methods (exclusion and inclusion) affect these ETFs’ financial performance. There is evidence that different socially responsible attributes might have different effects on financial performance. For example, Derwall et al. (2005) find firm-specific abnormal returns on environmentally clean firms, Edmans (2011) and Derwall et al. (2011) on firms with high employee satisfaction, and Bebchuck et al. (2009) on firms with good corporate governance, and Kempf and Osthoff (2007) on firms with good environmental performance. Humphrey and Tan (2014) argue that using exclusion can result in increased risk and lower returns. In contrast, Nofsinger and Varma (2014) show overperformance for funds that use positive screening in the ESG dimension. Similar to Nofsinger and Varma (2014), this thesis aims to examine the differences between screening strategies used (inclusion and exclusion) by SRI ETFs and what are attributes (ESG and product-related) that create the value in the SRI ETF data sample.

H2: Alphas vary between the different ESG attributes and screening strategies used in the SRI ETF sample group.

(12)

While there is much existing literature combining social responsibility and mutual funds, there is not much evidence on SRI ETFs’ performance. It is crucial to understand can these instruments offer a better alternative to mutual funds or other similar investing products. Since both the ETFs and social responsibility are gaining attention and attracting investors, it is essential to understand can these two phenomena create consistently better alpha in the passive asset management field. The socially responsible value is takin place in more extensive time periods, passive asset management and further ETFs offer a cost-effective way to practice this idea of doing good. Therefore, the study focuses on the last 11 years, offering the longest and most comprehensive research about SRI ETFs’ performance.

This thesis will offer answers for investors considering passive asset management over the active counterpart, and social responsibility as an investment philosophy, and ETFs as an instrument to practice these styles.

1.2 Structure of the thesis

This thesis consists of eight main chapters. The subchapter is an introduction of the topic, and it reasons why the topic is urgent for research. Additionally, the chapter argues what kind of question this thesis intends to fill.

To fully comprehend socially responsible ETFs’ performance, one has to understand the basics of these two phenomena. The second chapter aims to extensively review the theoretical background required to understand the asset management industry and the differences between active and passive asset management. The third chapter discusses the creating process of the ETFs and what kind of strategies they include. The third chapter also presents the history of ETFs and why they are gaining so much attention from investors. The primary reason for this chapter is to understand what kind of investment product the ETFs are.

(13)

The fourth chapter presents the theoretical framework of socially responsible investing.

Additionally, the fourth chapter introduces the market status of socially responsible investing and the strategies in this investing style. The fourth chapter’s primary purpose is to describe and discuss social responsibility as a potential to add or destroy the investor’s expected return of future investment value. The fifth chapter wraps up these two topics presented in the theoretical chapters of the ETFs and SRI. It presents the existing literature of these two phenomena as a combined investing product.

Furthermore, the chapter evaluates other studies and discusses the potential of these investment instruments.

After the literature review, the study moves on to the empirical part of this thesis. It presents the unique data set, how it is collected, and further what kind of methodology and empirical models are used to answer the research question. Chapter seven will present the results and analyze the results retrieved critically. Finally, chapter eight discusses the results and concludes the whole thesis.

(14)

2 Passive versus active asset management

This chapter describes the change in the asset management industry and how passively managed investment products are recently gaining popularity among investors, and further, why they offer an attractive alternative to the actively managed counterparts.

The following chapter also describes the underlying theories, such as the modern portfolio theory introduced by Markowitz in 1952 and the efficient market hypothesis by Fama in 1970. This chapter considers open-end funds like mutual funds and exchange- traded funds to be in the passive asset management category. Next chapter discusses the differences between exchange-traded funds and mutual funds.

Today’s investors enjoy countless investment products and vehicles to set their money into as an expectation of future profits. A wide range of actively managed portfolios and passively managed portfolios are available for investors to create their investment portfolios and achieve their investment objectives. Still, not forgetting traditional asset classes like common stocks, bonds, cash equivalents, and real estate (Bodie et al., 2014).

However, only seventy years ago, the asset management field was formed mainly of actively managed investment products, and no passive management products existed.

In his study, Sharpe (1963) suggests a shift to passively managed products.

The financial crisis of 2008 started an exponential shift in the asset management field.

The investors started to demand transparent investment vehicles with lower risk exposures as the crisis was occurred by risky and not investor-friendly alternatives like derivatives such as credit default swaps. Also, investors lacked confidence towards banks and financial institutions who generally are the asset managers. The focus from actively managed funds shifted towards passively managed funds measured in assets under management terms. The years 2005 and 2006 were the last two consecutive years that actively managed U.S. equity funds had back-to-back inflows. The assets under actively managed assets have grown but not like before. Actively managed U.S. equity funds have had outflows in 11 out of the last 12 years, while the passively managed U.S. equity funds have had extreme inflows after 2008. Figure 1 below presents that the passively

(15)

managed U.S. equity funds have caught up with the actively managed counterparts. Both were totaling at $4.3 trillion assets under management, reaching asset parity. The passively managed chart includes both mutual funds and exchange-traded funds.

(Morningstar, 2019a).

Figure 1. U.S. Equity fund flows between passive and active funds. (Morningstar, 2019a).

The origin of passive asset management can be linked to the capital asset pricing model that is emerged from the modern portfolio theory by Markowitz (1952) and the efficient market hypothesis by Fama (1970). One way to define passive asset management is to hold a diversified portfolio of assets without an effort on a security analysis to improve portfolio performance. Therefore, a passively managed portfolio or strategy has a minimal amount of additional input and relies on the benefits of diversification while trying to match the performance of a specific benchmark index. This kind of passive asset management strategy agrees that the financial markets are efficient and the prices contain all available information. (Fabozzi and Markowitz, 2011).

There are two kinds of passive asset management strategies: an indexing strategy and a simple buy-and-hold strategy. The indexing strategy is the more common one. The investors rely on a diversified portfolio of assets that follows a benchmark index’s

(16)

performance without attempts to identify any best performing stocks or emphasis of forecasting the stock market movements. Nonetheless, the buy-and-hold strategy involves elements of active asset management. In this strategy, the passive portfolio is purchased based on some stock selection criteria and then held for the predetermined investment period. (Fabozzi et al., 2011: 245-246).

The active asset managers assume the financial markets to be inefficient, where the active investors attempt to improve portfolio performance either by market timing or determining mispriced securities. An active asset management strategy exploits information asymmetry and uses forecasting techniques to achieve a better-expected performance than the passive investors, or the markets might achieve (Fabozzi et al., 2011). Sharpe (1991) points out how active asset managers as a group cannot beat the market. However, some managers beat the market. In the field of asset management, the overperformance results from either luck or skill of the active asset manager or investor. This skill relates to the active investor’s ability to obtain better information than most other investors, or the active investor can process the acquired information more efficiently to generate additional performance (Fabozzi et al., 2011: 247). Arnott and Darnell (2003) state that since investors invest in actively managed assets, they accept the value-added concept where the active asset manager can add value for investors through market inefficiencies and anomalies.

This chapter presented the extreme forms of passive and active asset management strategies. However, one can pursue many strategies, styles, and products as a passive or active investor in real life. The thesis’s interest is not to get in detail with them but rather to make a difference between extreme forms of active and passive asset management styles. To understand passive and active asset management products, one must understand the financial markets where these investment products are trading.

The underlying discussion about active versus passive asset management is closely related to the tests of efficient market hypothesis and how the markets accurately price the investment products.

(17)

2.1 Asset management under the efficient market hypothesis

The foundations of market efficiency were first introduced by Bachelier (1900), who studied the stock market’s behavior. Afterward, the idea of predicting stock market price movements and how the financial markets price investments were examined by Kendall (1953), who analyzed stock price behaviors over time. Kendall demonstrated that he could not identify patterns in stock price movement. This study gave basis to the discussion that stock prices follow a random walk and stock price movement is random and unpredictable. However, it was Fama in 1970 who made substantial developments for the efficient market hypothesis, which is considered to be the most influential study in the field of efficient markets (Bodie et al., 2014). In his article, Fama presents three forms of market efficiency. All of the forms in the efficient market hypothesis categorize the available information level in the financial markets. The three forms of the efficient market hypothesis are: “the weak form,” “the semi-strong form,” and “the strong form.”

The weakest form of the efficient market hypothesis (EMH) is “the weak form” of market efficiency, and it signals the lowest level of information in the markets. All historical information available in the markets determines the asset prices. This information includes historical trading prices, old interest rates, and past transaction volumes. Fama (1970) presents the old information as unimportant as the information is not a basis for future price fluctuations. Therefore, investors who use only available information about the past cannot generate additional profit since past information is not a guarantee for the future.

The intermediate form of market efficiency is the semi-strong form. In this form of the efficient market hypothesis, the asset prices are reflecting all of the released available information about the prospect asset. For example, the released information can include financial statements, balance sheets, dividends, annual wages, and current interest rates.

Fama (1970) states that in this form of market efficiency, prices reflect all the old information like in the weak form. Additionally, the current information available in the markets is reflected in prices as well. Once again, the investor cannot make any

(18)

additional profits based on the old information since stock prices will adjust again and again immediately to the news reveals that contain information regarding future stock prices.

The last and the strongest form of efficient market hypothesis suggests that all the current asset prices are set to a level where they include all available published information and all unreleased information. In this form of market efficiency, no party, like an investor or insider, has any “monopolistic access to price relevant information.”

Fama concludes that even if one can obtain inside information or unreleased information, the investor cannot benefit from that since the prices reflect that information already.

(Fama, 1970).

In the extreme form of the efficient market hypothesis, no one would actively analyze securities and commit resources to beat the market. Sharpe (1991) concludes that active asset management is always a zero-sum game, and after all the fees, it is a negative-sum game. To demonstrate, Sharpe considers the market portfolio to consists of passive and active investors. A passive investor holds a portfolio of securities that is the cap-weighted market portfolio. An active investor is anyone who does not hold a cap-weighted market portfolio. The aggregate market portfolio is the sum of passive and active investors, and therefore, all passive investors hold the market portfolio. The rest positions in the market consist of active investors, and as aggregate, they hold the same market portfolio. As a result, the passive investors earn according to the market portfolio before fees, and the active investors, in aggregate, earn the same market return before fees. Again, if the costs are higher for active investors, they lose to passive investors as aggregate. The only way to perform better than the market is to have skill or luck, and if a smarter active investor wins, it will be away from another active investor. Also, when taking the fees into account, the passive and active investors are both losing to the market. It can explain the increased concern in the investment field related fees and cost among investors discussed later. (Sharpe, 1991).

(19)

The actual level of market efficiency has been examined academically with conflicting results. Compared to Fama’s efficient market hypothesis spectrum, historical asset performance data implies that the markets are not even weak-form efficient. Jegadeesh and Titman (1993) show significant results towards weak-form market efficiency with momentum strategy where investors can obtain abnormal returns through analyzing past asset returns. In addition, the idiosyncratic volatility anomaly (Ang, Hodrick, Xing, and Zhang, 2006) implies that prior high volatility on an asset creates lower returns in the future. Contrarily, Chordia, Subrahmanyam, and Tong (2014) and Grobys and Haga (2016) support the strong form of market efficiency. They provide evidence that most anomalies exist only for a short period of time, and anomalies’ profits decrease after they have been discovered. Further, Fortin and Michelson (2005) show that not all markets are efficient.

Since the markets’ efficiency is hard to capture, Pedersen (2018) demonstrates that the markets are efficiently inefficient, where the markets are in equilibrium between the efficient and inefficient. Pedersen discusses why the arithmetic of active management introduced by Sharpe (1991) cannot exist in the real financial markets. Pedersen demonstrates that the active asset managers enhance the markets to be efficient, and the efficiently inefficient markets are equilibrium of passive and active investors.

Pedersen argues that since investors are paying for active asset managers in fees, the markets must be inefficient so that the active asset managers can outperform the market or the aggregate asset management field is inefficient since investors would pay fees for nothing. In addition, Pedersen points out that humans are players in financial markets, and humans are not always acting rationally. The asset managers and investors as humans can make errors, panic, herd or act in an irrational way that makes market prices fluctuate from the actual fundamental prices. Pedersen suggests that “the markets are just inefficient enough that active investors and their money managers can be compensated for their costs and risks through superior performance and just efficient enough that the rewards to money management after all costs do not encourage the

(20)

entry of new managers.” Therefore, the active asset managers provide liquidity to the markets and are compensated for providing service to investors. (Pedersen, 2015).

The efficient market hypothesis is fundamental for testing different academic models in the financial markets. However, the extreme forms are generally not accepted by asset managers and investors. An important outcome is that active and passive investors need to coexist in the market equilibrium. Both parties are essential players to make the markets more efficient. The active investors’ trade securities towards actual fundamental prices, and passive investors benefit from the low cost of indexing where short-term price fluctuations are not essential for their long-term performance.

2.2 Asset management under the modern portfolio theory

The underlying discussion between passive and active asset management is also affected by the portfolio creation process and allocation between different asset classes. In the modern financial markets, investors are spoiled with investment alternatives like presented before. This allocation between different assets is one of the most critical decisions the investor and asset managers do face when generating an optimal portfolio.

Markowitz (1952) introduces the modern portfolio theory where risk-averse investors create optimized portfolios by enchaining expected returns without increasing the risk of the portfolio. This can be called the risk-return tradeoff. The modern portfolio theory expresses the benefits of asset diversification in portfolio construction. (Bodie et al., 2014: 215-220).

In theory, investors are considered to behave rationally, and therefore maximize utility as in profits with the given level of accepted risk. Therefore, a risk-averse investor is assumed to choose a portfolio with the lowest risk if the available portfolios are yielding the same returns. Assets with higher risk should be compensated by higher returns, all else equal. Markowitz (1952) suggests that the investors try to minimize the portfolio variance for each expected gain of return, and vice versa, maximize the expected gain of

(21)

return for each level of variance, as variance meaning of risk. Thus, the decision-making when constructing an optimal portfolio is a choice between the mean and the variance of different assets. The outcome is that the portfolio with the highest expected return might not be the best on a risk-adjusted basis. Under these assumptions, all of the possible optimal portfolios existing in the financial markets construct the investor’s minimum-variance efficient frontier (Markowitz, 1952). Fama and French (2004: 26) discuss about the mean-variance model but refer to the same minimum-variance efficient frontier.

The minimum-variance efficient frontier presented below expresses all the portfolios formed by different risky assets. The risky assets mean that there is uncertainty involved in the outcome of investing in this asset. The frontier presents the optimal portfolio with the highest expected return for any level of risk. Identifying this frontier is the first step in portfolio management. All other portfolios on the frontier, below or above the global minimum variance portfolio, can be considered inefficient since investors can achieve a better risk-return tradeoff by choosing the global minimum-variance portfolio.

(Markowitz, 1952).

Figure 2. The minimum-variance frontier (Bodie et al. 2014: 220).

(22)

Tobin (1958) distinguished a weakness in Markowitz’s model. Investors can hold cash or other assets with very low risk. The modern portfolio theory is primarily based on a diversification strategy and is concerned with covariances between risky assets. The global minimum-variance portfolio is a combination of assets with returns that are less than perfectly positively correlated. As mentioned, there are many different assets for investors to choose from. Further, a risk-free asset is as an asset as well, for which the return is known with certainty today. In finance, a risk-free asset is commonly a short- term obligation of the U.S. Government like the T-bill (Bodie et al. 2014). The separation theorem suggests that if the investor has access to a risk-free asset like the T-bill, the optimal portfolio’s decision-making is absolute and independent of the investors’

preference for expected return and variance (Elton and Gruber, 1997).

The separation theorem has a couple of implications. Firstly, a tangency portfolio is formed by the risk-free asset when it creates a line through the expected return and standard deviation spectrum. This tangency portfolio is a portfolio that sorts the portfolio allocation problem by maximizing the ratio of the expected return minus the risk-free asset rate of return compared to the standard deviation. Secondly, this separation theorem constitutes to the optimal portfolio so that investors can achieve it by mixing two mutual funds. In this case, one of the funds made by the risk-free asset and the other one illustrating the tangency portfolio. Therefore, investors can achieve the global minimum-variance portfolio by holding two mutual funds. Elton et al. (1997) describe this as a mutual fund theorem. In addition, Elton et al. direct the future discussion about portfolio theory to the inclusion of debt and liabilities into the asset allocation decision-making process. The asset managers and investors using leverage need to account the borrowing rate into the portfolio management process. (Elton and Gruber, 1997).

The limitations of Markowitz’s (1952) modern portfolio theory are that the mean return distributions, the variance of assets, and correlations constantly move over time.

Modern portfolio theory calculates the optimal portfolio for a single time period only.

(23)

Therefore, as asset management is taking place over multiple time periods, the optimal portfolio at any given time is continually changing. Mossin (1968), Fama (1970), Hakansson (1974), and Merton (1990) examine this multiple time period problem left behind by Markowitz’s original work in 1952. The outcome of all the papers is that the multiple time period should be a sequence of single time periods. Hence, the optimal portfolio should differ from the one in a single time period (Elton et al., 1997).

However, essential for this thesis are the outcomes promoting passive asset management that have emerged from Markowitz’s work. Based on Markowitz’s modern portfolio theory, Bogle (1999) suggests that the most efficient and intelligent way to invest is using a diversified mutual fund to reduce the market risk. Samuelson (1965) suggests that investors would perform better by buying and holding a passively managed index fund rather than an actively managed fund or buying and selling individual assets.

Sharpe (1963) presents that investors cannot outperform the market consistently, and a passive index would most likely beat the active asset managers. Further, Malkiel (1973) suggests that the financial markets are efficient enough that most investors would perform better by allocating their wealth into a passively managed fund to mimic the market’s performance.

Finally, in 1975 Bogle established the first index fund called the Vanguard 500 index trust for passive investors to track the market performance. In the year the fund was established, it had $11 million assets under management, and by the year 1995, it had grown to $18 billion assets under management. Up to date (2020), the fund has close to

$600 billion assets under management. In his 2002 paper, Bogle demonstrates how passively managed outperforms the actively managed counterpart both on an absolute and risk-adjusted basis in 8 out of 9 categories during the ten-year sample period.

(24)

2.3 Performance of passive and active asset management

The question of whether investors should be passive or active is two-folded. Both participants are required to keep the markets at an equilibrium level of efficiency, and one cannot exist without the other. Besides, this question is extreme since different degrees of passive and active asset management exists. Academic researchers and asset managers disagree whether it is the passive or active asset management strategy that performs better than the other and is always subject to different personal preferences, data samples, and time-frames.

Some researchers (Arnott et al., 2003; Fortin & Michelson, 2002; Miller, 2007) have suggested that passive asset management provides more value given the higher fees and expense ratios of active asset management. Commonly, the investors pay fees to asset managers who manage the fund through load fees in the time of purchase or sell and operation expense ratios tied to the assets under management. Since active asset managers do trade securities more often than the passive counterparts, they tend to incur more expenses. Therefore, the active investors must weigh the expected return over the risks and cost of every trade. The general consensus in the academic literature is that lower fees are preferable to higher fees since higher fees destroy the expected returns (Gruber, 1996). Carhart (1997) suggests that expense ratios, portfolio turnover, and load fees affect the portfolio performance significantly and negatively. These results are consistent with Samuelson (1965), Sharpe (1966), and Malkiel (1995).

The Morningstar 2019 study of passive and active fund fees, presented below in Figure 3, demonstrates the difference in fees that investors pay for the asset managers regarding both styles and the overall decrease in fees after the year 2000. An investor using the actively managed funds did pay fees nearly four and a half times more than a passive fund investor during 2018. This study also suggests similar results to Carhart’s (1997) study. The low-cost funds are more likely to outperform the more expensive counterparts. The study by Morningstar has three important outcomes regarding asset management fees. Firstly, the investors are paying more awareness for the importance

(25)

of asset management fees which has led investors to favor funds with lower costs.

Secondly, the asset managers have realized the cost awareness of investors, and they have reacted to competition by cutting fees to vie for market share from the others in the asset management industry. Thirdly, the move towards lower cost and fee-based financial advice has aroused demand for lower-cost funds like exchange-traded funds.

(Morningstar, 2019).

Figure 3. Asset-weighted average fees for funds as the end of 2018. (Morningstar, 2019).

Since the active asset managers charge and incur more costs, it is only rational that the call for better performance is raised and expected to justify the higher fees. Carhart’s (1997) study offers slight evidence for skilled or informed asset managers who can beat the market. Carhart suggests that the performance after fees can be related to the one- year momentum effect by Jegadeesh and Titman (1993). Gruber (1996) presents that the cost factors can explain only some portion of the active asset managers’

underperformance.

As a result, Chevalier and Ellison (1999) questions whether some asset managers perform better because of personal characteristics and that some asset managers are therefore more skilled than others. They focus on the asset managers’ characteristics

(26)

rather than on the characteristics of the managed funds. They suggest that asset managers who have accomplished better selective undergraduate studies have better performance than asset managers with less selective undergraduate studies. In addition, they present that older asset managers perform worse than younger asset managers. To conclude, the findings suggest that some asset managers can perform better than other asset managers. This is paradoxical to the perfectly efficient market hypothesis, and the authors suggest that the difference in performance is only essential to keep equilibrium at the informationally efficient markets. (Chevalier et al., 1999).

Since few active asset managers can beat the market, studies like (Johnson and Collins, 2000) continue to call for active asset management. Their study suggests that only active asset managers can manage risk contrarily to passive asset managers. They highlight the active asset managers’ ability to shift assets to alternative investments and lock in profits when they arise. Similarly to the studies (Henriksson & Merton, 1981; Henriksson, 1984;

Andreu et al., 2018), who provide evidence of asset managers’ market timing ability.

Some studies like Grinblatt and Titman (1992) evidence the asset managers’ ability to collect information efficiently and pick the right stocks to achieve better performance consistently over time.

The S&P Dow Jones Indices report (2019) studies actively managed U.S. equity funds against their benchmark indices on a risk-adjusted basis over different time periods. To account for fees, they study the performance by including the fees as well as excluding the fees. They also adjust the benchmark returns by their volatility. After adjusting for the risk factors, the actively managed U.S. equity funds do underperform their passive benchmarks indices in all time frames and including the fees. Figure 4 below presents the net-of-fee underperformance percent by active funds compared to the benchmark indices. The study shows that even after controlling the fees, most active funds do underperform their benchmarks. This study is in line with Sharpe (1991), who highlights that, as a group, the asset managers cannot beat the market, and only very few managers can do it. (S&P Dow Jones Indices LLC, 2019).

(27)

Figure 4. Actively managed U.S. equity funds outperformed by benchmarks. (S&P Dow Jones Indices LLC, 2019).

Passive asset management is one of the most successful innovations of modern finance (Blitz, 2014). Investors are increasingly shifting from actively managed products to passive investment vehicles since no consistent overperformance for active management can be identified, and a surefire way to increase profits is to lower costs.

This shift is establishes on the discussion between the level of efficient markets (Fama, 1979) and modern portfolio theory (Markowitz, 1952) as well as their implications. The shift is also highly motivated by the studies on costs and performance differences. The start of passive asset management can be attached to the invention of the first index fund by Bogle (1975). However, passive asset management has always existed in the form of a buy-and-hold strategy. After the financial crisis of 2008, the investors’ demand for transparent and lower-cost investment vehicles grew explosively. Therefore, exchange-traded funds have received much attention and represent one of the most successful financial innovations during the last decades (Lettau and Madhavan, 2018).

(28)

3 Exchange-traded funds – ETFs

This subtitle describes the theoretical background of a comparatively new investment instrument; The Exchange Traded Funds (ETFs). In the interest of this thesis, it is essential to understand mechanics, strategy in creating processes, and the reasons for ETFs’

market share growth. As yet, it is crucial to understand the risks and costs of ETF investing.

McMahon (2005) presents the ETFs development as follows: “In the beginning, there were stocks; then came stock indexes, which offered professional assessment and instant access to multiple asset classes. Then came mutual funds, intended to lower the cost of exposure to the stock market and access multiple asset classes. Next came exchange-traded funds, hybrid security, intended to offer even better exposure, transparency, and economy to the investor.”

The American Stock Exchange (Amex) introduced the first ETF in 1993 called Standard and Poor Depositary Receipt (SPDR), which was afterward nicknamed as a “Spider.” The ETFs enable investors to trade entire index portfolios as they trade shares of stocks. The ETFs can is a basket of stocks that trade as a whole in an exchange. The “Spiders” were the beginning for many other similar products, designed to follow types of indexes such as DOW JONES, NASDAQ 100, or S&P500. The ETFs can follow a range of indexes from domestic to international or a more niche sector like a region or a country index (McMahon, 2005). At the end of 2019, there were around 5000 ETFs outstanding globally.

At the beginning of the 21^st century, ETF assets were fewer than 100 billion U.S. dollars and in 2019 counts over 6.18 trillion U.S. dollars. In addition, worldwide ETFs assets grew at an organic annualized rate of 18 percent from 2009 through 2019. ETFs were outpacing other open-end fund types by a growth rate of 4.8 percent. This is presented below in Figure 5. These numbers should be viewed over the whole $160 trillion asset universe consisting of global market value of equities and fixed income securities.

(Statista, 2020; Lettau and Madhavan, 2018).

(29)

Figure 5. Assets under management of global ETFs from 2003 to 2019 (Statista, 2020).

3.1 Mechanics and characteristics of ETFs

The reasons for ETFs’ market growth and popularity are many. Most ETFs seek to track a specific index just like a mutual fund does. This tracking process is indexing, which expresses the replication strategy that the ETF follows the target benchmarks assets in all market situations. However, there are also actively managed ETFs in the markets that try to outperform their benchmark index, but most of the ETFs still follow a target benchmark. Due to their index mimicking strategies, the ETFs are for long-term passive asset management. The ETFs are generally low-cost, transparent, liquid, and tax-efficient and offer easy diversification for investor’s portfolios. The ETFs are relatively simple and easy to use since their similarity to common shares. They can be bought and sold in exchanges intraday, and investors can monitor prices in real-time. (Lettau and Madhavan, 2018).

(30)

Dellva (2001) critics ETFs for their intraday trading and that flexible trading rules create an environment where investors chase short-term capital gains chasing a hot sector or fund. Investors use ETFs’ trading features, although these instruments are for long-term investors for them to match a specific index’s performance. However, Barber and Odean (2000) provide evidence that investors trading more are prone to lower returns compared to returns from investors who trade less.

Today ETFs offer a wide range of alternatives as well. Diversification and the number of alternatives lead to an easy and extensive way for investors’ risk management. Like it turns out, ETFs have characteristics from both mutual funds and common company shares. However, since they offer benefits from both, investors do not have to pick any specific share or fund but to decide the area of the markets to invest in. For example, diversification provides a good option when investing abroad. Therefore, the optimal portfolio by (Markowitz, 1952) can be achieved by combining different ETFs that maintain benefits from the different correlations between assets. The ETFs are also a convenient and cost-effective way for small investors to reach special markets that would be too expensive or otherwise complicated to access. Thus, ETFs offer new opportunities for investors like a piece of a share that would usually be too expensive to own. (Delfeld, 2007: 1-2; Lettau & Madhavan, 2018.)

Understanding ETFs mechanics better, it is easy to compare them with conventional mutual funds. A mutual fund holds the underlying assets. For example, an S&P 500 index fund holds a portfolio of shares that make up the S&P 500 index. Over time, the mutual fund manager takes the responsibility to maintain the portfolio this way, and if an investor redeems from the fund, the mutual fund manager needs to adjust the underlying portfolio by selling assets. In turn, the ETF investors operate in the secondary markets through an exchange or a broker and other liquidity providers. The intra-day transactions between investors on the secondary markets do not cause transactions in the underlying assets to which the asset manager needs to react. Therefore, the

(31)

mechanism does not lead to any transaction costs like in the mutual funds when adjusting the portfolio in case an investor redeems. (Lettau et al., 2018).

In addition, more cost benefits for ETFs arise over mutual funds when the mutual funds interact directly with the investors. Typically mutual funds incur distribution and record- keeping costs such as transfer agency costs as well as different services fees ranging from marketing to distribution that the ETFs do not face. The ETFs offer investors also more transparency since the holdings are listed daily when the mutual fund holdings are listed quarterly. The ETFs also incur tax advantages through the “in-kind” transactions that reduce capital gain distributions for investors. The “in-kind” process will be discussed later in this chapter. (Lettau et al., 2018).

The Securities and Exchange Commission (SEC) requires the ETF and mutual fund managers to publish a net asset value (NAV) for their funds. The ETFs operate contrary to mutual funds, whose sales occur only once a day when the fund’s new net asset value is determined based on the component securities’ last recorded quotations. So in a mutual fund, all transactions occur at the end of each trading day and at net asset value, when ETFs are traded throughout the trading day at their net asset value. The net asset value is calculated as the total value of the funds’ underlying assets (the value of holdings in cash, shares, bonds, derivatives, and other securities) minus the total value of its liabilities and fees. The net asset value is then again divided with the total shares outstanding to determine each fund’s share price. The net asset value formula is presented below. Both the ETFs and mutual funds NAV derives from this formula. (Lettau et al., 2018).

𝑁𝑒𝑡 𝐴𝑠𝑠𝑒𝑡 𝑉𝑎𝑙𝑢𝑒 (𝑁𝐴𝑉) = 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑢𝑛𝑑−𝑇𝑜𝑡𝑎𝑙 𝑙𝑖𝑎𝑏𝑙𝑖𝑙𝑖𝑡𝑖𝑒𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑢𝑛𝑑

𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠ℎ𝑎𝑟𝑒𝑠 𝑜𝑢𝑡𝑠𝑡𝑎𝑛𝑑𝑖𝑛𝑔 (1)

For international mutual funds and ETFs, the net asset value can be adjusted to take into account the market movements in other markets since ETFs can be listed to the U.S., but they hold securities from other markets. For example, an ETF trading in the U.S. that

(32)

holds an asset from the Tokyo exchange is valued for the closing price in Japan but adjusted for changes in the yen/dollar currency rates during the U.S. trading day. (Lettau et al., 2018.)

A significant role in the ETF markets play the so-called authorized participants. In contrast to a mutual fund, ETFs do not interact with the markets directly. ETF asset manager, like Vanguard, BlackRock or State Street enters into a contract with these authorized participants, generally large financial institution, who in return interacts with the markets and investors. The authorized participants are responsible for controlling the ETFs in the markets. They are the ones who create or redeem ETF shares that are to say, authorized participants act as dealers for the ETF shares and control, for example, the ETF liquidity. A large ETF can have 38 authorized participants as an average to minimize the risk that one would finish their activities. Creation of all of the new ETFs and current ETFs extinguished through these processes called creation and redemption.

This process is the key mechanism to control the price changes and hold an ETF price as close to the target index as possible. This mechanism presented here above is called the

“in-kind” mechanism and is demonstrated below in Figure 6. (Lettau & Madhavan, 2018.)

Figure 6. The ETF architecture (Lettau & Madhavan, 2018).

(33)

The ETFs trade in two distinct markets. The primary market is where the authorized participants and the ETF manager create and redeem ETF shares for underlying securities, which occurs at the net asset value of the ETF. The secondary market is the exchange where the investors sell and buy the listed ETFs. The redemption and creation process helps to keep the actual market value of the ETF and the net asset value in line with each other. The ETF intraday trading behavior might differ, and the ETF can be trading at a premium or discount to the fundamental value. This distinction from the actual value can be profitable for fast-movers who can distinguish such opportunities in time. In case the price is below the net asset value, the ETF is trading at a discount. The fast-mover can purchase ETF shares, redeem them for the underlying assets, and again sell the underlying assets to their actual market prices, which will end up the ETFs net asset value.

In case the price is above the net asset value, a fast mover can do the reverse and create new ETF shares based on the underlying assets. These can also be referred to as arbitrage activities. For example, the authorized participant may trade ETFs when they estimate that security is trading at a premium, and typically, the authorized participant will lock any profit intraday. (Lettau et al., 2018).

Through arbitrage activities, ETFs increase informational efficiency in the markets by decreasing the short-term mispricing between net asset value and the market prices.

Glosten, Nallareddy, and Zou (2016) demonstrate how ETFs increase the short-run informational efficiency of the ETF underlying assets, especially for smaller firms with less analyst coverage. Petäjistö (2017) provides evidence that ETFs holding liquid U.S.

domestic assets are priced efficiently. In contrast, the ETFs with illiquid or international assets are economically significantly deviating with 100-200 basis points from their actual prices and therefore suggest mispricing to remain in some inefficient markets.

Due to their liquidity benefits, Itzhak et al. (2018) argue that the high-frequency investors trying to benefit from the mispricing can increase the volatility of the ETFs, which is not the desired effect.

(34)

Engle and Sarkar (2006) examine the premium and discount of U.S. domestic and international ETFs from a day-to-day and minute-by-minute perspective. Their model investigates time variations of the standard deviations of these ETF pricing errors. Their result suggests that the domestic U.S. ETFs have only a small (15 basis point standard deviation on average) and highly transient pricing errors lasting only a couple of minutes, while the international ETFs have larger and longer pricing errors that can last several days. They explain this by higher transactions cost of the creation and redemption process for international ETFs in addition to wider bid-ask spreads.

The premium or discount occurs when the ETF manager is not able to track the benchmark index correctly. Therefore, the NAV returns vary from the benchmark index returns the ETF is supposed to track. The unexpected spread between the price of underlying securities in the ETF portfolio and the benchmark index price is called as a tracking error. The following equation is called the “NAV tracking error,” and it is defined as in Tang and Xu (2013) and Piccotti (2018) as:

𝑁𝐴𝑉 𝑇𝑟𝑎𝑐𝑘𝑖𝑛𝑔 𝐸𝑟𝑟𝑜𝑟 (𝑇𝐸_𝑖,𝑡) = (𝑟 _𝑖,𝑡^𝑁𝐴𝑉− 𝑟 _𝑖,𝑡^{𝐼𝑛𝑑𝑒𝑥}) × 100, (2)

where 𝑟 _𝑖,𝑡^𝑁𝐴𝑉 is the daily arithmetic NAV return of the ETF, and the 𝑟 _𝑖,𝑡^{𝐼𝑛𝑑𝑒𝑥} is the daily arithmetic return of the targeted benchmark index. Therefore, the NAV Tracking Error measure is positive if the ETF net asset value outperforms the index, and vice versa, negative when the benchmark index outperforms the ETF.

Piccotti (2018) suggests that ETFs usually trade above their net asset value on a premium.

He argues that the ETF investors are willing to pay a premium to achieve liquidity benefits that the ETFs can provide for investors, for example, when granting exposure indirectly to inaccessible underlying securities. These inaccessible underlying securities can be foreign equities, fixed income assets, or anything the investor does not have direct access to due to high-cost accessibility or location reasons.

(35)

Gastineau (2010: 69-72) argue that some of the tracking error involved in the ETFs can be due to Regulated Investment Companies (RICs) and Undertakings for Collective investment in Transferrable Securities (UCITS) regulations. Both of these regulations are aiming to harmonize the financial markets in their continents. The U.S. Internal Revenue Code (IRC) determines specific diversification requirements for regulated investment companies to enable favorable tax-free treatment. The ETF portfolio holdings and their magnitude can restrict the tax-free distribution of interest, dividends, and capital gains to shareholders. The RIC regulation obligates that no more than 25 percent of the ETFs assets can be assets of a single issuer except the U.S. government. In addition, now in line with 50 percent of the ETFs total assets, no more than 5 percent of the assets can be from the same issuer and again expect from the U.S. government. Therefore, the minimum number of required assets is 13 as with the following allocation, two assets with no more than 24.9 percent each, and ten assets with no more than 4.9 percent each, and one asset holding the remaining 1.2 percent of the total ETF portfolio.

In Europe, the Undertakings for Collective investment in Transferrable Securities (UCITS) regulation is a bit more complex and is not regulation for tax-free treatment. This regulation is also known as the 4/10/40 rule. The ETFs under UCITS can invest no more than ten percent of total assets in transferable assets or money market securities of the same issuer. Now respect to 40 percent of total assets, the ETFs can invest no more than five percent of total assets in the same issuer’s securities. Therefore, under the UCITS regulation minimum number of required assets is 15. Table 1 below presents this allocation of assets under the RIC and UCITS regulation. Gastineau (2010) notes that even with the regulations, many ETFs could meet both of these regulations’

requirements relatively easily, but for many ETFs, they might cause replication hazards and therefore cause tracking error. (Gastineau, 2010: 69-72).

(36)

Table 1. Minimum diversification requirements for RICs and UCITS. (Gastineau, 2010: 71- 72).

3.2 Strategies used in forming an ETF

The extreme forms of passive and active asset management strategies are in the last chapter, and both of these can be implied to ETFs. The ETFs can have dimensions of both where the asset manager follows a specific benchmark or, on the other hand, tries to enhance the performance of the ETF compared to the benchmark. This is only a lavish breakdown. As ETFs have attracted a lot of fund inflows recently and therefore grown significantly in assets under management, diversity, and market significance, new types of ETFs have started to emerge. In addition to ETFs that follow a specific index, ETFs can be categorized to domestic and international equity ETFs, bond or fixed-income, commodity, currency, real estate, or a wide range of different ETFs that exploit, for example, a specific style like the momentum or social responsibility. However, the ETFs that track a specific index are still dominating the industry. (Lettau et al, 2018; BlackRock, 2020).

(37)

The domestic or international equity ETFs typically also track an index or benchmark of equities. The equity ETFs can cover a wide range of stocks domestically or from different countries internationally. They can also cover large or small business stocks or target sector stocks like pharmaceutical stocks, tech stocks, or banking stocks. The bond or fixed income ETFs invest in different types of bond markets, like the U.S. corporate bond markets, government bond markets, emerging bonds markets, or green bond markets.

The commodity ETFs track the price changes of a specified commodity like gold, silver, or oil. The commodity ETFs usually own derivates to mimic the performance rather than owning actual commodities because storage costs for physical commodities would drive the ETF costs up. Further, there can be commodity equity ETFs that invest in stocks of commodity producers. The currency ETFs hereby invest in currencies like the U.S. dollar or Japanese yen. They either own a single currency or a basket of currencies. The real estate ETFs or real estate investment trust (REIT) ETFs invest in a type of real estate, or in a broader perspective, the real estate markets. The factor ETFs invest through a rule- based approach that targets specific return divers across all asset classes. These specific drivers can be a metric like a dividend growth, expected volatility, or high-growth potentiality. Eventually crucial for this thesis are the different styles that can be incorporated into the ETF strategy decision making like the socially responsible ETFs.

(BlackRock, 2020).

Additionally, ETFs can be categorized as physical or synthetic ETFs, which means what kind of replication strategy ETFs use to form the index they try to follow. In any case, the objective of all ETFs is to follow the performance of the benchmark index with a minimal tracking error. Physical ETFs try closely replicate an index by holding the underlying securities. These securities are exchanging between the ETF manager and authorized participant. These two allow the creation and redemption of units in response to shifting demand from investors. In other words, ETFs replicate the underlying index physically investing in the underlying securities and in the same ratio as in the index. Physical ETFs are hence very transparent between the ETF and the benchmark index. However, criticism against physical ETFs is the transaction costs due to continuous trading on

(38)

securities and that they carry more tracking error when comparing to synthetic ETFs.

(Maurer and Williams, 2015).

Synthetic ETFs do not hold the underlying securities in the index. Instead, they perform thorough derivatives, most commonly through swaps. The synthetic replication method is more commonly taking place in Europe because of the Undertakings for Collective investment in Transferrable Securities (UCITS) regulation (Gastineau, 2010). Successfully executed securities lending can provide significant returns and meanwhile maintain a low-risk profile. A key advantage of a synthetic ETF is that it should track an underlying index much more exact than a physical ETF due to its nature with derivatives. As a result, synthetic ETFs should carry lower tracking error risk, but they are facing counterparty risk due to securities lending. In practice, the ETF manager and the authorized participant form a swap agreement that commits to yield a particular index’s return without owning the securities themselves. In return, the ETF manager pays cash to the counterparty. In addition, there are synthetic ETFs that use multiple counterparties to provide flexibility to change exposure if there were concerns about counterparties’

creditability. This is done by using different parties as in ETF manager as well as in the authorized participant. However, Maurer and Williams (2015) study shows that the physical ETF follows the benchmark index with similar efficiency to synthetic ETFs, and the investor is not benefiting from carrying the counterparty risk. (Maurer and Williams, 2015).

Besides physical and synthetic ETFs, there are relatively new products called the leveraged and inverse ETFs or alternatively, called bullish and bearish ETFs. These are more active investing strategies, and both are designed to seek more performance than the benchmark depending on the market cycle. The leveraged ETFs aim to beat the underlying benchmark and deliver typically twice or three times the benchmark performance over a specific period, typically on a daily basis. The inverse ETFs aim to short the market and to deliver performance opposite to the underlying benchmark also on a daily basis. Inverse leveraged ETFs, on the other hand, try to short the underlying

Value creation with passive socially responsible exchange-traded funds